Nvidia is approaching a breakout level. Plus, what we want to see from Micron

August PCE inflation rose 3.4% year over year, down from 3.7% in July, while core inflation undershot expectations, pushing Treasury yields lower and supporting equities. Nvidia has gained 15% over the past three months and is nearing its $236.54 May record high, with CoreWeave planning deployments of Nvidia's Vera Rubin platform. Micron rose ahead of earnings as AI-driven memory demand continues to exceed supply, while Boeing initially gained 2% after winning the Pentagon's Navy sixth-generation fighter-jet contract before reversing lower.

Analysis

The lower-rate impulse supports long-duration AI equities, but the more important distinction is between companies with validated demand and those relying on externally financed capacity buildouts. NVDA can benefit from technical momentum above its prior high, yet CRWV deployment announcements should be treated as demand-quality evidence only after confirming contracted capacity, funding sources, and utilization; circular vendor/customer financing remains the key multiple-risk across the AI infrastructure complex. A failed NVDA breakout while yields continue falling would be a useful warning that AI positioning, rather than rates, is the marginal constraint.

MU earnings are the highest-quality near-term catalyst in this set because the debate is shifting from current pricing strength to the duration of HBM and high-end DRAM tightness. Upside requires evidence that supply discipline extends beyond the next two quarters and that customer agreements protect pricing into 2027; otherwise, the market will discount a conventional memory-cycle peak despite strong reported results. Boise execution is strategically valuable but not an immediate earnings driver, while any delay would reinforce concerns over U.S. fab cost inflation and subsidy dependence.

LLY's combination approach increases the probability that obesity becomes a broader, longer-duration treatment market rather than a zero-sum share fight. That is structurally negative for NVO if superior efficacy is matched by tolerability and manufacturing scale, though payer reimbursement—not trial efficacy—is the 6-18 month constraint. BA's defense award should not materially alter the equity thesis absent disclosure on development funding, fixed-price exposure, and production economics; investors should avoid extrapolating a headline win into near-term free-cash-flow improvement.

Key Decisions for Investors

• Maintain a tactical long NVDA through the next 1-3 months only if it closes decisively above the prior peak on expanding volume; use a failed breakout and relative underperformance versus SOXX as a stop signal. Prefer NVDA over CRWV: the latter has greater refinancing and customer-concentration sensitivity if rates or AI utilization assumptions reverse.
• Buy MU only into earnings with defined downside via a call spread or reduced cash position rather than chasing pre-print strength. Upside case is management extending constrained-supply commentary beyond two quarters; exit if gross-margin guidance implies normalization sooner than consensus expects.
• Establish a 6-12 month relative-value position long LLY / short NVO in equal dollar exposure after reviewing durability and discontinuation data. The trade works if efficacy differentiation broadens the addressable population and payer coverage; it is falsified by tolerability limits, supply bottlenecks, or NVO matching the clinical profile.
• Do not add to BA on the defense-program headline. Reassess only after contract structure and cash-flow timing are disclosed; fixed-price development losses or renewed commercial production disruption would outweigh the strategic value of the award.